Fig. 1 The influence of sample water potential on the observed slope of the temperature correction factor given in Eq. [1]. (A) relationships for in situ stem psychrometers (PWSI) and (B) for C-52 sample chambers (Wescor). In both panels open symbols show results for a single salt standard placed in each psychrometer followed by a series of temperature changes without ever opening the sample chambers, while closed symbols indicate values derived from full calibrations performed at each temperature. n = 3 psychrometers per model per water potential for continuous readings, and 6, and 9 psychrometers for Wescor and in situ stem, respectively, during the multiple calibration "reload" method. Error bars indicate one standard error. The plotted lines are functions chosen for their empirical fit. A single regression was fit to the data from both calibration methods employed: (A) temperature sensitivity slope (i.e., Eq. [1]) = (-2.10 x 10-4)
2 - (2.606 x10-3) + 2.333 x 10-2 (r2 = 0.97) and (B) temperature sensitivity slope = [(3.511 x 10-2)2]/[2 + 4.269 x 10-2)] (r2 = 0.97). Note that all regressions are scaled to give a value of 1 for Ø at 25°C, so the respective y intercepts are always implicit in the slopes
